Popular
Share this on social media:
Paweł Malinowski, program manager at Imec
Imec’s Paweł Malinowski looks at the SWIR imaging landscape, as ST Microelectronics announces a quantum dot SWIR sensor at IEDM
Acquiring information in the shortwave infrared (SWIR) wavelength range has been for decades limited to niche applications. SWIR is used in the military, to identify targets in adverse lighting conditions, for instance; in machine vision, such as in solar cell inspection; and in scientific and space systems. These are typically high-end applications with a cumulative annual number of units sold in the tens of thousands globally.
SWIR cameras are traditionally based on focal plane arrays with a resolution rarely exceeding 1 megapixel, and a pixel size an order of magnitude larger than what’s found in CMOS image sensors (CIS) used in consumer devices – a typical pitch in a SWIR detector is 15-20µm versus sub-micron pixels for CIS. Imagers are made by hybrid bonding detector chips based on III-V materials – usually InGaAs, sometimes HgCdTe – to the readout chip using solder bump connections and die-to-die flip-chip techniques. All those elements in the fabrication process result in a very high price point for these sensors, easily exceeding several thousand euros per chip.
A recent innovation reported by Sony in 2019 uses Cu-Cu bonding of an InGaAs detector chip, which resulted in scaling down the pixel pitch to 5µm. The IMX990 and IMX991 chips are finding their way into more and more products, especially in machine vision, as they enable not only higher image quality but also easier integration. Even with a premium price point, this sensor family opens up SWIR imaging to new use cases.
A technology that promises to lower the implementation barrier even further are SWIR sensors based on thin films, with quantum dots playing the lead role as a new type of absorber. The technology has been investigated for almost two decades; the first academic papers were published at the beginning of the 21st century by Professor Sargent’s group at the University of Toronto, with pioneering work since then.
Now, the first products are being brought to market. Emberion recently announced its VS20 camera with a broad spectral range, from 400 to 2,000nm. Meanwhile, SWIR Vision Systems offers its Acuros series of cameras, with the highest SWIR resolution on the market (1,920 x 1,080 pixels). These are disruptive products in the infrared imaging field and prove that the maturity of quantum dot technology is sufficient to deploy commercial products.
In parallel, research centres such as Imec continue to explore the QD pixel stacks, readout architectures, and the integration process to improve performance and enable upscaled fabrication. At the same time, new application fields are opening up for end-users that didn’t consider SWIR imaging before because it was simply beyond their reach, in terms of cost but also size and complexity.
In 2019, Imec presented a 5µm pixel pitch QD SWIR image sensor, and in 2020 topped that with a 1.82µm proof-of-concept device – both state-of-the-art pixel density for SWIR image sensors at the time of introduction.
Now, yesterday, at the 67th International Electron Devices Meeting (IEDM), the tone-setting event for the semiconductor industry, ST Microelectronics announced QD image sensors. The 1.62µm pixel pitch sets a new record, and the external quantum efficiency of 60 per cent at 1,400nm inches ever closer to the values found in the incumbent technologies.
The most exciting feature is that these results come from chips fabricated using a 300mm wafer platform – this means that the QD technology has made significant strides in making it ready to manufacture. Upscaling to a wafer-level process promises extraordinary throughput and thus cost evolution. New sensor products based on this process will disrupt the market further, and act as a critical enabler for SWIR imaging in applications that never considered it before, including consumer products.
Imaging in the SWIR range offers features such as improved eye safety for devices using laser light like lidar – eye sensitivity to radiation at wavelengths above 1,400nm compared to 940nm is around six orders of magnitude lower – as well as for low-light imaging and cameras that can see through adverse weather conditions.
Looking forward, there are still technological challenges to address. Moving away from lead sulphide-based QDs to lead-free material systems will encourage more players to accept this type of sensor. Improving deposition throughput of the quantum dot layers by going to one-step coating – instead of the standard layer-by-layer coating used currently to achieve the desired final thickness of the absorber – will significantly increase the takt time in volume production.
Moreover, thorough investigation of reliability metrics according to industry standards will set further iterations of QD improvement to optimise long-term stability, and enable even the most harsh applications such as those demanded by the automotive sector.
ST Microelectronics’ announcement at IEDM should reassure the industry that quantum dot image sensors are a credible technology with excellent potential that can be upscaled to manufacturing in semiconductor foundries. This should fuel further investment in research of these fascinating devices, and spark a new wave of applications for more accessible SWIR camera systems.
—
Paweł Malinowski is program manager at Belgian R&D institute Imec. He has more than10 years experience in developing photonics technologies, and is currently working in the thin-film electronics group of Imec.
The first colour image, captured by a CMV20000 sensor, to be sent back by the hazard cameras on the Perseverance rover after its landing on Mars on 18 February. Credit: NASA/JPL-Caltech
Images taken with ST Microelectronics' 940nm NIR Quantum Film sensor (top left) and with its 1,400nm SWIR QF sensor (bottom left). Corresponding images taken using a visible smartphone camera (right). The QF NIR image shows better contrast between black electrical wires hidden in the dark green leaves, and tree trunks and branches hidden in front of the dark wood fence. The SWIR QF image shows how effective it is to use SWIR imaging to see through a silicon wafer. Credit: ST Microelectronics
Image: Valery Lisin/shutterstock.com
Image: Martial Red/shutterstock.com
Image: Konstantin Zibert/shutterstock.com
Greg Blackman finds out how seed producers are reaping the rewards of imaging
Image: Denis Simonov/shutterstock.com
Tahir Rabbani of Viztronics spoke at the Vision show about a 3D solution for imaging shiny orthopaedic implants. Greg Blackman reports
Image: asharkyu/shutterstock.com
Greg Blackman speaks to Pleora’s James Falconer about the latest versions of GigE Vision
Image: Sonpichit Salangsing/shutterstock.com
As solar power becomes an increasingly important energy source, Abigail Williams looks at how infrared imaging is used to assess solar panels
Image: shutterstock.com
Anne Wendel, director of VDMA Machine Vision, says there is increasing concern about chip supply bottlenecks and material shortages
Image: Landemesse Stuttgart GmbH
Huge strides are being made in neural networks, but there are still significant challenges to implement a machine vision solution, according to experts at Vision
Image: Sony Semiconductor and Prophesee
Prophesee has won the Vision Award for its event-based approach to imaging. Greg Blackman speaks to Luca Verre, the firm’s CEO, about prospects for the technology